摘要
将数据挖掘技术应用于入侵检测中可以提高检测的精度和效率。针对k-means算法对初始聚类中心很敏感，在聚类过程中对数据输入的顺序也有依赖性等特性，本文首先利用遗传算法初始中心点对k-means聚类算法进行了改进，然后使用k-means算法快速收敛获取聚类结果，最后在入侵检测的经典数据集KDDCUP1999上检验了算法的有效性。实验结果表明，该方法与相关研究对比提高了入侵检测系统的精度和效率。
The data mining technology applied to intrusion detection can enhance inspection accuracy and efficiency.Because K-means algorithm of initial clustering center is very sensitiveit is also depend on the order of data input in the process of clustering,then this paper ftrstly use the initial center of the genetic algorithm for the improvement of k-means clustering algorithm,and use the fast convergence features of k-means algorithm for obtaining clustering results,and finally test the effectiven -ess of the algorithm in intrusion detection of classical data sets （KDD CUP 1999）.Experimental results show that compared with the intrusion detection system,the method and relevant study improve the efficiency and precision.
出处
《计算机光盘软件与应用》
2011年第3期67-68,92,共3页
Computer CD Software and Application
